Carbon monoxide adsorption on magnesium oxide - The Journal of

Adsorption Equilibria and Kinetics of Carbon Monoxide on Zeolite 5A, 13X, MOF-5, and ... Electron-Donating Property of Sodium Metal-Doped Magnesium Ox...
0 downloads 0 Views 715KB Size
Carbon Monoxide Adsorption on Magnesium Oxide

1019

Carbon Morio idle Adsorption on Magnesium Oxide

. St. (2.

7. L. Slager,lb L. H. Little,*'a and R. G. Greenlerlb

School of Chemistry, The University of Western Australia, Nedlands, Western Australia, 6009 and Laboratory for Surface Studies and Department of Physics, University of Wisconsin-Milwaukee,

Milwaukee, Wisconsin 53201 (Received October 12, 1972)

Downloaded by UNIV OF WINNIPEG on August 27, 2015 | http://pubs.acs.org Publication Date: April 1, 1973 | doi: 10.1021/j100627a011

The adsorption of carbon monoxide onto magnesium oxide was investigated. The fQrmation of surface carbonate groups was observed on samples prepared in high vacuum while added oxygen was necessary for carbonate formation with magnesium oxide subject to more vigorous outgassing in ultrahigh vacuum. Carbon dioxide produced similar carbonate species to that formed by carbon monoxide.

Introduction Several spectroscopic investigations of the adsorption of carbon monoxide on oxide surfaces have shown that surface carbonate-type species are formed on the surface.2 In many instances the surface species produced are identical with those obtained in the adsorption of carbon dioxide. In order to investigate the possible origin of the additional oxygen needed to ccnvert carbon monoxide to adsorbed carbonate radicals an infrared spectroscopic study has been made of the aldsorption of carbon monoxide onto magnesium oxide sarnples prepared under different conditions of evacuation. I[n one set of experiments, adsorption studies were made on samples prepared in a system capable of producing 10-5 Torr, while in another an ultrahigh vacuum system producing 10-*-10-9 Torr was employed. In the latter experiment it was found that carbon monoxide did not produce the typical spectra of adsorbed carbonatle surface species unless additional oxygen was supplied to the system. The adsorption of carbon dioxide on magnesium oxide was used in the work as a basis for comparison of the species produced by adsorbed carbon monoxide. Evans and Whateley" have recorded infrared spectra from 4000 to 700 cm-1 of various coordinated carbonate species from carbon dioxide adsorbed on magnesium oxide at temperatures in the range 20-500". They have also investigated the in.teractioru of adsorbed carbon dioxide with water and deuterium oxide. The spectrum of carbon dioxide adsorbed at room temperature is complex and at present is not fully understood. It is apparent, however, that the method of preparation of the surface, influencing the degree of dehyciroxyylation of the surface, has a considerable effect on its adsorptive properties. The studies of Evans and WhatePey were carried out on a surface prepared at 850°, corresponding to a substantially dehydroxylated surface. The work described here shows the effect on the infrared spectrum of carbon dioxide adsorbed on samples not subjected to extreme dehydroxylation. xperimental Beetion

Torr Experiments. Magnesium oxide was obtained by decomposition at 600" in air for 3 hr of precipitated magnesium hydroxide prepared from AR reagents. Pellets of the oxide were prepared and immediately placed in the infrared cell which was evacuated promptly t o minimize adsorption of atmoapheric carbon dioxide. The infrared cell employed in this study was described by Cant and L i t t l ~ Before .~ adsorption was commenced the sample was heated in oxygen foip 0.5 hr at 400" and evacuated at 450" for 2.5 hr a t 10-5 Torr. Under this treatment the sample

surface was substantially hydroxylated as shown in its spectrum. Pressures between 20-and 200 Torr qf carbon monoxide and carbon dioxide were employed. All spectra were recorded using a Perkin-Elmer 521 spectrophotometer. The cell used in the adsorption studies was connected directly to a vacuum line enabling all evacuation, heating, and adsorption processes to be carried out in the infrared beam. A similar sample, pretreated in the same manner as the adsorbent, was set up in the reference beam of the spectrometer. Both cells were connected to the vacuum line and evacuated similarly. Carbon monoxide or carbon dioxide was admitted to the cell in the sample beam and an equal pressure of helium was admitted to the reference cell to balance any difference in the thermal emission from the compensating sample. Exchange of the surface hydroxyl groups was effected by admitting deuterium oxide for 0.5 hr at 80" then evacuating at 300" for 1 hr. A second admission of D2Q was made to achieve more complete exchange. lo-8-10-9 Torr Experiments. The infrared spectra in this section of the work were recorded with a Beckman IR9 spectrometer under conditions giving a resolution of 1 cm-I. The vacuum system was constructed of stainless steel using a 1.5-in. disk of Irtran I1 (Eastman Kodak) as infrared windows. These windows were attached to the vacuum system as described by Kottke and Greenler5 with GeVac sealant obtained from General Electric. The pressure was monitored with a diaphram gauge, a thermocouple gauge, and an ion gauge. The composition of residual gas in the system a t a total pressure of 1 P 5 Torr or below was determined with a Veeco MMIT mass spectrometer. The sample was heated by radiation from a wire enclosed in an evacuated quartz envelope constructed of two concentric quartz cylinders, 48 and 58 rnm in diameter. This arrangement eliminated contamination of the sample from the tungsten wire filament. The magnesium oxide was obtained from Electronic Space Products, Inc., and was quoted at 4 N purity. Carbon monoxide, oxygen, and carbon dioxide were obtained from Matheson as analyzed research grade a&a purity of 99.9% or better. The water used was distilled water which had been degassed prior to use. (a) School of Chemistry, The University of Western Australia. (b) Laboratory for Surface Studies and Department of Physics, University of Wisconsin. Reviewed by L. H. Little, "Infrared Spectra of Adsorbed Species," Academic Press, London, 1966, p 74. J. V. Evans and T. L. Whateley, Trans. Fafaday SOC., 63, 2769 (1967). N. W. Cant and L. H. Little, Can. J. Chem., 48, 1373 (1968). The Journal of Physical Chemistry, Vol. 77, No. ti, 1973

Smart, Slager, Little, and Greenler

1020

A sample of magnesium oxide was pressed at 4000 lb/ in.2 and fired for 1 hr a t 950" in air prior to placing it in the vacuum system The sample was then heated to 500". A pressure of 8 >: 10-8 Torr was recorded while at 500". The major residuals were H2, " 2 0 , CO, and (202. Upon cooling to room temperature the pressure dropped to 5 >: 10-9 Torr. Either carbon monoxide or carbon dioxide was then added and spectra were run a t various time intervals.

40

9

Y

W V

230 I-

C

Downloaded by UNIV OF WINNIPEG on August 27, 2015 | http://pubs.acs.org Publication Date: April 1, 1973 | doi: 10.1021/j100627a011

I

Results Adsorption of Carbon Monoxide Torr Euacuation). The adsorption of carbon monoxide (100 Torr) a t room temperature on a magnesium oxide surface prepared by 10-5 Torr evacuation at 450" is shown in the spectra of Figure 1. The absorption bands are similar to those observed for carbon dioxide adsorption (Figure 5) with an additional band discernable a t 1595 cm-I. However the extent of the initial rapid adsorption is not as great nor is the subsequent rate of growth of the bands as rapid as that observed for carbon dioxide adsorption. The intensity of carbon monoxide bands is approximately one-fifth that of the carbon dioxide bands a t similar pressures. Similar changes in the spectrum of the surface hydroxyl groups (or deuterioxyl groups) on the magnesium oxide were observed with increasing adsorption of carbon monoxide to those observed during carbon dioxide adsorption. Deuterium exchamge of the hydroxyls to deuterioxyl groups improved the quality of the spectra since the displacement to lower frequencies moved the bands to a spectral region where the scattering losses were less severe. The sharp Y OD band at 2766 cm-l from isolated OD groups decreased in intensity with increasing adsorption while the broad band of perturbed OD groups a t 2750-2100 cm-1 intensified (Figure 7B) as the surface concentration of adsorbed carbonate radicals increased with more COz adsorption. Similar effects were observed during 6 0 adsorption except that the gaseous C02 band at 2330 cm-1 was absent. The addition of gaseous oxygen (100 Torr) to the system produced changes in the relative band intensities (Figure 1F). The 1660- and 3 3 9 0 - ~ m -bands ~ increased in intensi-

Figure 1. (A) Spectrum of magnesium oxide. (5) 100 Torr of carbon monoxide admitted for 10 min, (C) for 6 hr, (D) for 24 h r , (E) for 72 h r , (F) 100 Torr of oxygen added to carbon monoxide in t h e cell. Spectrum recorded 24 h r after oxygen admission. The Journai of Physical Chemistry, Voi. 77, N o . 8, 7973

tn z

U

E20

I

:

?so0

1600

1400 1200 WAVENUMBER (CM")

1000

Figure 2. Adsorption of CO on MgO in a leaking system: (A) background and initial CO spectrum; (B) 2.4 hr later; (C) 5.8 hr iater; (D) 69.7 hr later. ty more rapidly than the other bands. By comparing the growth of bands during adsorption and the disappearance of bands during desorption studies, it was possible to assign pairs of bands to individual surface species. The band pairs are as follows: 1520 and 1400 em-l, 1660 and 1390 cm-1, and 1690 and 1310 cm-I. Similar pairs of bands were found for adsorbed CO2. An additional pair a t 1595 and approximately 1350 cm-1 were observed during CO adsorption. Similar spectra for adsorbed carbon monoxide on magnesium oxide were reported by Kolbel, et a1.6 CO Adsorption i n UHV System. One of the early experiments with this system produced the spectra (Figure 2) showing the very rapid growth by bands in the region 1700-1000 cm -I. When evacuation was attempted a large leak was found to have developed during the experiment. No changes were detected immediately after addition of the carbon monoxide shown as trace A. The spectrum a t 2.4 hr (B) showed three major bands a t 1675, 1307, and 980 cm-1 and a minor band a t 1400 cm-l. Further standing (C) produced major changes seen as a loss or serious shift of the bands a t 1675 and 1307 em-l to give peaks a t 1650, 1510, and 1375 cm-l. The 1 4 0 0 - ~ m -absorbance ~ was either lost under the 1375-cm-l peak or shifted to this frequency. Standing for another 64 hr (D) yielded a further loss of detail and a new band at 1270 cm-f. A fresh sample was placed in the repaired system and surprisingly no change was observed from the background spectrum upon addition of 24 Torr of carbon monoxide to the system. Trace A of Figure 3 represents both the background and magnesium oxide i n a carbon monoxide atmosphere after 22 hr. Addition of 26 Torr of oxygen produced a n immediate change in the spectrum (Figure 3B) and standing for 6 hr (Figure 3C) amplified these new peaks. They appear at 1670, 1307, 1275, and 980 em-1. Further standing for 69 hr (Figure 3D) produced a shoulder at 1375 cm-1 and a broadening of the 1670-em-1 band to lower frequency. Spectrum A of Figure 4 is a repeat of D in Figure 3 to allow the investigation of the effect of evacuation on the spectrum. Pumping at 10- Torr for 2 days (B of Figure 4) resulted in a decrease a t 1275 and 980 cm-1 and the shoulders at 1535, 1375, and 1055 cm-1 be(5) M. Kottke and R. G. Greenler, Rev. Sci. Instrum., 42, 1235 (1971). (6) H. Kolbel, M. Ralek, and P. Jiru, Z. Naturforsch. A , 25, 670 (7970)

Carbon Monoxide Adsorption on Magnesium Oxide I

1800

1800

I

1400 1200 W'AVE NU M BE R (C M"')

1021 I

I

1

1000

Downloaded by UNIV OF WINNIPEG on August 27, 2015 | http://pubs.acs.org Publication Date: April 1, 1973 | doi: 10.1021/j100627a011

2o

1 1409

I600

1000

I200 R '

WAVENUMBER ( C M 1

'Figure 3. Adsorption of CO on MgO: (A) background and 22 h r after 24 Torr of CO addition; (B) addition of 26 torr 0 2 ; (C) 6 h r after B; (D) 74.7 hr after B.

-1400 1200 WAVENUMBER (CM-'1

BOr

1000

Figure 4. Evacuation and further reaction of the sample of Figu r e 2: (A) 74.7 h r after addition of 0 2 ; (B) evacuation to 2 X 1O-' Torr; (C)addition of 1 Torr of H20.

came more defined. 'The subsequent addition of 1 Torr of water to the system (C of Figure 4) produced three major bands a t 1650, 1500, snd 1390 cm-l. CO7 Adsorption ( 1 0 - 5 Torr System). The adsorption of CQz (20 Torr) a t room temperature on a magnesium oxide surface prepared by 10-5 Torr evacuation at 450", is shown in the spectra of Figure 5. Initial adsorption produced bands at 2330, 1680, 1520, and 1390 cm-f (with a shoulder a t 1330 cni-1). The growth of the bands with time shows that the 1520- and 1 4 0 0 - ~ m -bands ~ increase most rapidly, particularly after heating the sample in the presence of 6 0 2 ilt 150". Transmission through the sample in the region 1100-800 em-1 was poor although bands formed during the adsorption of CO2 were detected a t 1000-900 cm 1 and a t approximately 850 cm- I. Desorption studies showed that the 2 3 3 0 - ~ m -band ~ was removed immecliately on evacuation a t room temperregion ature. The remaining bands in the 1800-1200-~m-~ were removed in re'verse order of their appearance on adsorptxon. Thub the species producing bands at 1520 and a t 1400 em-f i s only removed by evacuation for several hours at 4 " and is clearly the most stable carbonate species on the surface.

Figure 5. (A) 2 cm of carbon dioxide admitted to magnesium oxide for 0.5 hr; (B) 2 cm of carbon dioxide admitted for 12 hr; (C) the sample in B heated to 150" in the presence of the carbon dioxide; (D) background spectrum of magnesium oxide.

5

I800

'p

I

I

1600

I

I

1400

I

I

1200

1000

Figure 6. Adsorption of GO2 on MgO: (A) background; (B) after addition of 27 Torr of Con;(C) 26 h r later and evacuation to 2 X lO-'Torr.

C02 Adsorption in UHV System. 'The adsorption of 27 Torr of carbon dioxide on the magnesium oxide sample produced immediate and large changes from the background (A of Figure 6) at 1650, 1375, 1307, 1275, 1055, and 980 cm-1 (B of Figure 6). A very broad shoulder was also detected at approximately 1535 cm-1. Standing for 26 hr and pumping to 2 X 10-7 Torr (C of Figure 6) gave a broadening in the 1650-1500-cm-1 region and at 1375 cm-f and a decrease in intensity a t 1275 cm-l. Heat Treatment and Evacuation. When the carbon monoxide or the carbon dioxide experiment was terminated prior to the addition of water, heating to 500" and evacuation to 8 X 10-8 Torr reproduced the background. During this evacuation procedure the maximum pressure was detected between 200 and 300". A mass spectrum taken during the pressure maximum showed that COz was the major component discharged from the surface whether CO or C02 was admitted initially.

Discussion The results show that uncontrolled leaking of atmosphere into the UHV system could be duplicated by conThe Journal of Physical Chemistry, Voi. 77, No. 8. 1973

Smart, Slager, Little, and Greenler

Downloaded by UNIV OF WINNIPEG on August 27, 2015 | http://pubs.acs.org Publication Date: April 1, 1973 | doi: 10.1021/j100627a011

1022

trolled addition of oxygen (Figure 3B) with the addition of water vapor as discussed below. A surprising feature was the complete absence of bands detected in the infrared spectrum (Figure 3A) when only carbon monoxide was placed in the system. With the addition of oxygen, absorption bands are produced in the spectral region characteristic of carbonates. However, adsorption of carbon monoxide proceeds readily to produce surface carbonate species in the system evacuated a t 10-5 Torr. It has not been determined whether the additional oxygen needed for carbonate formation is supplied by the residual gas atmosphere in this system or by oxygen or some weakly held oxide species in the adsorbed state which is not removed in the less severe evacuation conditions of this system. Assignment of Bands to Surface Carbonates. Characteristic vibrational band assignments in the infrared spectra of various carbonate complexes have been reviewed by Little.2 In brief, bands a t 1520 and 1370 cm-l, together with bands in the region 1060 and 850 cm-I, can be assigned to a unidentate carbonate ligand. Bands at 1630 and 1270 cm -1 with others a t 1030 and 830 cm-l may be assigned to bidentate carbonate ligands. The spectra (Figure 3B) obtained when oxygen and carbon monoxide are admitted to magnesium oxide suggest that two bidentate surface ligands are formed with bands a t 1670, 1307, alnd 980 cm-I and 1670, 1275, and 980 cm-1. The broadness of the 1670-cm-l band (Figure 3B) suggests that two o#verlappingbands occur here, although no broadening or splitting of the 980-cm-1 band is apparent. These bands arise from the carbon-oxygen stretching modes of the adsorbed bidentate species. An additional form of adsorbed carbonate species is produced on standing oxygen and carbon monoxide over magnesium oxide or on heating the sample. Shoulders appear at 1535, 1375, and a t 1055 cm-1 (Figure 3D). These bands are also prodiiced when COz is adsorbed on magnesium oxide and can best be assigned to unidentate carbonate groups. The two bands a t high frequency arise from the asymmetric and symmetric stretching modes of the two uncoordinated carbon-oxygen bonds of the unidentate carbonate ligand and the 1055-cm-l band from the stretching of the bond through which the carbonate is Coordinated to the surface magnesium ion. Desorption studies a t increasing temperatures show that the stability of the carbonate species follows the order unidentate ~ > bidentate (1670- and (1520- and 1 3 9 0 - ~ n i - band) 1307-cm-I band) :> bidentate (1670- and 1275-cm-l band). No explanation can be given for this order. This order of stability was also found by Kolbel, et aL6 A comparison with the spectra of other magnesium compounds7 shovvs MgC204~2HzO to be a bidentate structure having bands a t 1675, 1380, 1310, 1020, and 815 cm-1, while MgC(38 is a simple carbonate with bands at 1450, 1010, 890, and 750 cm-l. Thus the surface carbonates formed are noi, the simple carbonate of the normal magncsium carbonate structure, but have bidentate structure !similar to MgC204.2Hz0. The pressure increase in the system when the adsorbed carbonates are heated shows that these species decompose between 200 and 300" i n uacuo. MagiTesiuLm carbonate decomposes a t 365" in air.8 Since the major deccmposition product is COz, as detected during the pressure maximum, the carbon monoxide and oxygen must form a carbonate, which subsequently decomposes to magxle3ium oxide and C02. The addition of waiter (Figure 4C) produced a structure with very similar spectra to that of MgC03.Mg(OH)2. The Journal of Physical Chemistry, Val. 77, No. 8, 1973

L 4wo

3000

zoo0

w AVE NUMBER

1580

1000

( c rn"' )

Figure 7 (A) ( i ) Magnesium oxide spectrum: ( i i ) after exchange of surface hydroxyl groups with deuterium oxide. (E) (i) Magnesium oxide spectrum after deuterium oxide exchange as in A; (ii) after admission of 2 cm of carbon dioxide for 10 min; ( i i i ) After admission of 2 cm of carbon dioxide for 64 hr.

3Hz0.7 The separate stepwise addition of oxygen and water duplicated the observations of the leaking system. A very large number of spectra recorded both during adsorption and desorption on the 10-5-Torr vacuum system show that the two carbonate stretching modes a t higher frequency (1700-1250 em-1) are grouped in pairs about the frequency (1450 cm-1) of the isolated carbonate ion. Thus a band of highest frequency, 1670 cm-1, will be associated with one a t lowest frequency, 1270 cm-l, while other pairs will be intermediate in spacing. The band a t 2330 cm-l ( Figure 7 ) is due to weakly coordinated or physically adsorbed carbon dioxide and i t is readily removed by room temperature evacuation.

Surface Hydroxyls and Interaction with Adsorbed Carbonates. Figure 7 shows the spectrum of a magnesium oxide pellet prepared by lO--g-Torr evacuation a t 450". Malinowski, et al., 9 have investigated magnesium oxide prepared at various temperatures in uucuo. Their spectrum in the hydroxyl-stretching region is similar to that in Figure 7 with a sharp and two broad bands. The sharp band a t 3756 cm-1 is due to free, noninteracting hydroxyl groups, while the broad bands a t 3690 and 3590 em-1 indicate hydroxyls involved in weak hydrogen bonding. It is possible that the broad band at 3590 cm-1 is due at least in part to internal hydroxyl groups. 'This i s supported by results obtained with deuterium exchange of the hydroxyl groups. After exchange the free hydroxyl band disappears completely and the corresponding free OD band appears a t 2766 em-1 (Figure 7). Two types of bonded deuterioxyl groups are also indicated by bands a t 2700 and 2650 cm-1. However, a broad band of medium intensity remains a t 3580 cm-1 after exchange. This may be attributed, as Davydov, Kiselev and Zhuravlevlo have shown for silica surfaces, to internal hydroxyl groups which are more difficult to exchange. It i s possible that some exchange of internal (7) The Sadtler Standard Spectra, Nos. 1732, 1740, and 1763. (8) "The Handbook of Chemistry and Physics," 48th ed, Chemical Rubber Publishing Go., Cleveland, Ohio. (9) St. Malinowski, S. Szczepanska, A. Bieianski, and J. S!oczynski, J. Catal., 4,324 (1965). (10) V. Y . Davydov, A. V. Kiselev, and L. T. Zhuravlev, Trans. Faraday SOC.,60,2254 (1964).

Downloaded by UNIV OF WINNIPEG on August 27, 2015 | http://pubs.acs.org Publication Date: April 1, 1973 | doi: 10.1021/j100627a011

Radicals Bonded to Porous Vycor Glass

1023

hydroxyl groups will take place a t 80" when hydrogen- and deuterium-containing species could diffuse into the bulk of the sample. Tf, ixi fact, the band a t 3580 cm-1 in the spectrum before exchange is due to internal hydroxyl groups then the band at 2650 cm-I after exchange is probably due to some internal deuterioxyl groups. Evans and Whateley3 have assigned a broad absorption found after water admission in the spectrum below 1200 cm-1 to the deformation of surface hydroxyl groups. Hunt, Wisherd, and Eionhamll found broad absorption in this region with magnesium oxide prepared a t room temperature. This absorption, doubtless, contributes to the rising spectrum observed in the present study at frequencies below 1280 cm- l. The interaction between surface hydroxyl (or deuterioxyl groups) and adsorbed carbonate ligands can be seen in Figure 7B. Gas-phase carbon dioxide has overtone bands in the region 3720-3500 em-' which overlap the hydroxyl bands 2f the adsorbent. For this reason, as well as to reduce radiation scattering losses, the interaction of C02 with deuterioxyf groups was observed. With increasing adsorption of surface carbonate groups it was found that the sharp free OD band ait 2766 em-I decreased in intensity. A corresponding intensity increase in the bands due to bonded, deuteriowyl groups was observed (Figure 7B) in the 2700-2100-~:1-1-' region.

The interaction between adsorbed carbonate species and adsorbed water or surface hydroxyls may occur through hydrogen-bonding interaction as in I or through the production of a n adsorbed bicarbonate species as in II.

I1

I

The relatively small deuterioxyl (or hydroxyl) perturbations (Figure 7B) observed are in accord with the formation of a relatively weak hydrogen bond as in 1. Acknowledgment. We gratefully acknowledge grants from the Australian Research Grants Commission to support the work a t the University of Western Australia and from the National Science Foundation, to support the work at the University of Wisconsin-Milwaukee. (11) J. P. Hunt, M. 1478 (1950).

P. Wisherd,

and L. C . Bonham, Anal. Chem., 22,

decl to Porous Vycor Glass elamud,

M. G . Reisner, and U. Garbatski*

Department of Chemistry, Technion-krael

lnstitute of Technology, Haifa, lsrael

(Received October 16, 7972)

Publica tion costs assisted by the Department of Chemistry, Technion, l. 1. T.

In porous Vycor glass surface OH groups were substituted by methoxy, ethoxy, n-butoxy, tert-butoxy, and acetoxy groups. Various radicals containing a -cH2 group were obtained by ultraviolet irradiation and identified by their esr spectra. In most cases the radicals were stable even a t room temperature for days and weeks. This stability is probably due Lo chemical bonding. CHs radicals were also observed when methylated and tert-butylated glasses were irradiated a t low temperatures. Various possibilities for the mechanism of radical formation are discussed. An important aspect is the fact that the surface concentrations of radicals even at low temperatures and during irradiation are around lO14/m2, four orders of ma,:nitude below those of their parent compounds.

Introduction

A number of papers concerning free radicals stabilized on sohid surfaces have a ~ p e a r e d . l -It ~ seems that this phenomenon of small organic radicals stabilized up to room temperature is restricted to the methyl radical. The exceptional stability of this radical physically adsorbed on the surface is not yet fully understood. Other stable radicals originally thought to be physically adsorbed were subsequently found to be chemically bonded to the surface. Thus, Vladimirova, et aL,8 assigned the esr spectrum obtainLed by irradiation of methanol adsorbed in silica and alumina t o the CH20H radical. Ono

and Keiig found that the signal belonged to the -OCH:! group chemically bonded to the surface. Chemical reaction of alcohols with various surfaces has been mentioned by several ~ o r k e r s . l O - ~ ~ (1) ( a ) V. 6. Kazanskii and G . 6. Pariiskii, Fiz, Tverd. Tela, 5 , 473 (1963); (b) Zh. Strukt. Khim., 4, 336 (1963); ( c ) Prepr. Pap. Int. Symp. Free Radicals, 6th, 1963 (1963); i d ) Proc. lnt. Congr. Catal., 3rd, 1, 367 (1964). (2) C.L. Gardner and E. J. Casey, Can. J. Chem., 46,207 (1966). ( 3 ) J. Turkevich and Y . Fujita, Science, 152, 1619 (1966). (4) M. Fujirnoto, H. D. Gesser, B. Garbutt, and A . Cohen, Science, 154,381 (1966). (5) M. Fujirnoto, H. D. Gesser, B. Garbutt, and M. Shimizu, Science, 156, 1105 (1967). The Journal of Physical Chemistry, Voi. 77, No. 8, 1973